As one of the major male malignant tumors in urology, prostate cancer has gradually surpassed other tumors in incidence in recent years (1). Meanwhile, diversified treatments have also appeared in the management of PCa. Assigning the most suitable treatment for patients depends on the risk stratification of prostate cancer itself. As described in the latest EAU and NCCN guidelines, prostate cancer is now divided into low, intermediate and high-risk groups, or even more with intermediate risk group being subdivided into unfavorable intermediate-risk group and unfavorable intermediate-risk group. The basis of this classification mainly derived from the acknowledged risk factors related to the prognosis of prostate cancer, for example, preoperative PSA level, Gleason score composition (primary and secondary pattern), clinical and pathological staging (4, 21–23). GS plays an important role in risk classification so as in treatment strategy making. To those patients with non-regional and non-distant metastases prostate cancer, if GS ≤ 6 (low-risk), active surveillance or surgical treatment are both available while if GS ≥ 8 (high-risk localized), multimodal treatments are preferred, for example, radical prostatectomy or radical radiotherapy as initial treatment, androgen deprivation therapy (or radiotherapy) as adjuvant therapy. If GS = 7, either 3 + 4 or 4 + 3, radical prostatectomy or radiotherapy is recommended. However, one should receive adjuvant therapy after RP depends on whether he has lymph node metastases or adverse features, which now include positive surgical margin, seminal vesicle invasion and extracapsular extension (see NCCN). Identifying the risk factors for PCa patients with GS 7 is extremely important because it helps identify the certain group of patients who need adjuvant therapy and can benefit from it maximumly in prolonging survival times, meanwhile helps avoid over treatment and unnecessary side effects for those who don’t need adjuvant therapy.
GS was first proposed by Gleason and defined as the sum of primary and secondary patterns of prostate specimens (2, 3), which indicated that prognosis of prostate cancer was mainly decided by its primary and secondary Gleason patterns. Primary Gleason pattern was found to play an even more important role in predicting malignancy degree of PCa with more investigation being conducted. However, researchers were wondering what role tertiary but higher Gleason pattern played in PCa since GS was proposed until Pan C.C., et al. first proposed TGP was also a risk factor to PCa (5).
The main purpose of this study was to decide whether TGP5 was an independent risk factor to PCa with GS 7 in Chinese population and whether clinician should take it into account as adverse feature when assigning further treatment for PCa patients with GS 7 after RP.
There have been many studies published concerning TGP’s effects on prostate cancer. Pan, C. C., et al. first brought tertiary Gleason pattern to the public. They analyzed 114 radical prostatectomies with tertiary components (༜5%), which were compared with a prostatectomy database comprised of 2,276 cases without a tertiary component. In their study, GS 7 with TGP5 showed significantly worse pathologic stages than GS 7 without TGP5 and were not different statistically from typical GS 8 (4 + 4) tumors. GS 7 with TGP5 revealed significantly higher progression rates than GS 7 without TGP5 (5). Adam, M., et al. compared 2,396 patients with (22.4%) and 8,260 without (77.5%) a tertiary Gleason pattern for adverse histopathological features (extra-prostatic extension, seminal vesicle invasion, positive surgical margins and lymph node invasion) and analyzed the effect of a tertiary Gleason pattern on biochemical recurrence. They found TGP was statistically significantly associated with all evaluated histopathological parameters and it was an independent predictor of biochemical recurrence (HR 1.43, p < 0.001). On subanalysis, TGP independently predicted biochemical recurrence in patients with GS 3 + 4 and 4 + 3 after RP, respectively (6). Borhan, W. et al. analyzed a total of 4060 specimens with a GS 7 with and without TGP5. Cases were subdivided into 3 + 4, 3 + 4 with TGP5, 4 + 3, and 4 + 3 with TGP5. They compared prostate-specific antigen, clinical stage, pathologic stage, and surgical margin status between the groups. The impact of TGP5 on biochemical recurrence was also assessed. They found TGP5 was related to higher PSA level, pathologic stage, positive surgical margin and worse prognostic outcome (8). Although many studies got the same results that TGP5 related to one or more clinicopathological characteristics and PCa outcome (4, 21–23), some studies failed. It remains uncertain that what variables can predict TGP5 and whether 3 + 4 + TGP5 has same prognostic outcome as 4 + 3 without TGP5.
In this study, we retrospectively analyzed 229 patients who met with inclusion criteria and had complete and reliable data from Jan. 2014 to Dec. 2018 in Peking University First Hospital. Clinical and pathological characteristics distribution between different groups with and without TGP5 was being analyzed. Differing from some researches abroad, our study found that TGP5 wasn’t related to elder age at diagnosis (≥ 65 years old) and higher preoperative PSA level (≥ 10 ng/ml). Besides, no statistical difference was found between GS 7 without TGP5 and GS 7 with TGP5 when comparing prostate volume, PSAD, GS variation, clinical T staging, pathological T staging, T staging variation, extra-prostatic extension, seminal vesicle invasion and operation ways.
Recently, a study by Jiakun Li et al. from China was published on OncoTargets and Therapy. In their study, they included 350 patients in total from 2009 to 2017, among which ~ 10% (n = 34) had TGP5. They also found that TGP5 wasn’t related to elder age at diagnosis, higher preoperative PSA levels and pathological staging (24). In our study, ~ 25.3% (n = 58) patients had TGP5, which was higher than Jiakun Li’s study and also higher than some studies from other countries. Cases were all within 6 years, strictly followed the recommended criteria proposed during ISUP 2014 on TGP5, which meant our study were of more time-effectiveness and reliability. So, even we were surprised by the finding that TGP5 was not related to those variables, we could not fully rule out the possibility that TGP5 was indeed not associated with those variables mentioned above. We believed race and genetics difference might account for and explain it. But given that our sample size was still small, more investigation needs to be done in the future.
To be noted, the PSM rate in the entire cohort was 21.4% and we also found the presence of TGP5 was related to higher PSM with statistical significance (P = 0.002). In GS = 7 with TGP5, PSM rate was 36.2% while in GS = 7 without TGP5, PSM rate was only 16.4%. In Sauter G, et al.’s study, the PSM rates were 16.09%, 13.18% and 48.54% in GS = 7, GS = 7 without and with TGP5, respectively (25). In Li, J. et al.’s study, the PSM rates were 20.17%, 19.81% and 23.53% in GS = 7, GS = 7 without and with TGP5, respectively (24). Some other research also had similar findings that the presence of TGP5 was related to higher PSM rate while there weren’t many explanation or discussion about it (8, 16). Compared to these studies with larger population, our comparison results were consistent with them. However, PSM rate was higher in our cohort. Especially for group GS = 4 + 3 with TGP5, PSM rate was even higher as 48.1%. There were many factors including pathological T staging, operation ways, experience of surgeons etc. having impact on PSM rates, it was hard to attribute higher PSM to TGP5 alone. In Zhang Z. et al.’s study, they included 113 PCa patients with GS 7 after laparoscopic RP in Peking University First Hospital from 2016 to 2017 and the PSM rate of PCa patients with GS 7 was 30.97%. The number of positive cores, positive percentage of needle biopsy, and pathological stage were correlated with PSM rate (P < 0.05) in their study (26). In our view, TGP5 was related to poorer pathological characteristics, like higher T staging, according to other studies mentioned above (although it was not shown in our study), so PSM occurred more easily when TGP5 was present. But in our cohort, small sample size might be the source of deviation. Further investigation with larger population should be done to validate this.
Biochemical recurrence was the primary endpoint in our study, which is now considered as a sign of prostate cancer relapse. To identify the predictors of BCR is crucial since it will help to select patients needing multimodal therapy while spare the others. In our study, for prostate cancer with GS 7, the incidence of BCR in the group with TGP5 was significantly higher than that in the group without TGP5, indicating TGP5 was a risk factor to PCa. Kaplan-Maier survival analysis and log-rank test validated it with biochemical recurrence-free survival (BFS) curve showing that the presence of TGP5 was negatively correlated with BFS. Univariable cox-regression analysis got the same result that TGP5 was a risk factor to prostate cancer with GS = 7, HR 2.640 (95%CI 1.143–6.097). In univariable cox-regression analysis, GS composition and pathological T staging were also risk factors to PCa prognosis. Given that the composition of GS and pathological T staging might interfere with TGP5 in survival analysis, multivariable cox-regression analysis was also conducted, showing TGP5 was an independent risk factor to BCR, with HR 2.360 (95%CI 1.019–5.464).
This study had many limitations. Firstly, like other retrospective studies, recall bias was the biggest source of bias in this study. Secondly, sample size was small compared to previous studies, with only 229 cases finally entered into our analysis and only n = 23(10%) patients experiencing BCR. However, our cohort was of more time-effectiveness and reliability since the pathological reports of specimen accorded with the latest criteria from 2014 ISUP and variables affecting the accuracy of survival analysis were considered as exclusion criteria. Thirdly, there were some variables not recorded such as pelvic lymph nodes dissection (PLND) during operation. In this cohort, there were only 67 cases (29.26%) experiencing PLND during RP, and all were negative findings. According to the existing studies comparing oncological and non-oncological outcomes between no PLND and PLND, the benefits and harms of PLND during RP still remain controversial (27). Urologists in this hospital also have different opinions on whether performing PLND during RP or not. We finally decided to discard this variable to ensure the accuracy of the survival analysis. But we believed that lymph nodes dissection during RP was theoretically related to the survival of prostate cancer and might affect some results of this study.
We conducted a stratified analysis considering and eliminating the influence of GS composition on prostate cancer. The cohort was divided into 4 groups as 3 + 4 without TGP5, 3 + 4 with TGP5, 4 + 3 without TGP5 and 4 + 3 + TGP5. Results showed that overall differences among 4 groups were found in terms of GS variation, pathological T staging, prostate extracapsular invasion, positive surgical margins, and biochemical recurrence rate. But these differences mainly existed between group 3 + 4 without TGP5 and group 4 + 3 with TGP5. In 3 + 4/4 + 3 group alone in our study, the existence of TGP5 was not related to poorer clinicopathological features than 3 + 4/4 + 3 without TGP5. K-M survival analysis among 4 groups also showed statistical difference only between group 3 + 4 without TGP5 and group 4 + 3 with TGP5. BFS in group 3 + 4 with TGP5 were not statistically different from group 4 + 3 without TGP5. Whether it indicated that 3 + 4 with TGP5 had same risk as 4 + 3 without TGP5 was worthy of more investigations. Although the stratifying analyses did not find many positive findings or consistent with previous studies, due to the small sample size, this study was still the first stratifying analyses grouped by GS and TGP5 in China, which was the novelty of this study.
While exploring the prognostic effect of TGP on prostate cancer, researchers are also working on how to integrate tertiary Gleason pattern (5) into new grading standard in prostate cancer (with GS 7). Sauter, G., et al. proposed the integrated quantitative Gleason score (IQ-Gleason) to evaluate the risk of prostate cancer. It ranges from 0-117.5 and is calculated as follows: percentage of unfavorable Gleason pattern (Gleason 4 + Gleason 5) + 10 score points if any Gleason 5 pattern was seen + another 7.5 score points in case of Gleason 5 quantities of > 20%. For example, the IQ-Gleason of a Gleason 3 + 4 = 7 cancer with 40% Gleason 4 is 40, the IQ-Gleason of a Gleason 3 + 4 = 7/tertiary grade (TG) 5 cancer with 40% Gleason 4 and 5% Gleason 5 is 40 + 5 + 10 = 55. The IQ-Gleason of a (Gleason 4 + 5 = 9) cancer with 60% Gleason 4 and 40% Gleason 5 is 60 + 40 + 10 + 7.5 = 117.5. This model was published on European Urology and got some initial and promising results (25). Its quantification idea originated from Gleason Score but got more. Other models were also created by different researchers but lack of large population-based validation for now. These models remind that urological pathologists should keep an idea of quantitation when processing and reporting specimens after RP. We are looking forward to establish a new grading system for PCa concerning TGP to assign best treatment for patients.